Koalas in the Wild

Data Collection Process

The images in this dataset were scraped from the iNaturalist platform, which is widely used by citizen scientists and wildlife enthusiasts to document and share observations of species around the world. Each image in this dataset represents unique sightings of koalas in the wild, offering diverse perspectives from different states of Australia.

To ensure high-quality annotations, the dataset includes manual labeling of the x and y coordinates for each koala in the images. These annotations provide bounding box coordinates, offering a precise localization of the koalas within each image. The annotations are stored in a CSV file that contains the minimum and maximum values of the x and y-axis for every image.

Dataset Structure

The dataset consists of:

• Images: 2,100 images of koalas from different Australian states, each with varying backgrounds and settings, depicting koalas in the wild.
• Annotations: A CSV file with manually annotated bounding boxes, specifying the minimum and maximum x and y coordinates for each koala in the corresponding image.

Each row in the CSV file corresponds to a unique image and includes:

• Image ID: The identifier for the image.
• Bounding Box Coordinates: Minimum and maximum x and y values, outlining the koala’s location in the image.

Use Cases

This dataset can be used in a variety of applications:

• Wildlife Conservation: Assisting conservationists in tracking koala populations and understanding their movement patterns.
• Machine Learning & Computer Vision: Training object detection models to identify koalas in the wild, even in complex environments.
• Ecological Research: Analyzing the effects of environmental changes, particularly bushfires, on koala habitats and behavior.